A type of generalization error induced by initialization in deep neural networks.
Yaoyu ZhangZhi-Qin John XuTao LuoZheng MaPublished in: MSML (2020)
Keyphrases
- generalization error
- neural network
- model selection
- cross validation
- learning algorithm
- binary classification
- active learning
- training data
- target function
- sample complexity
- sample size
- upper bound
- classification error
- training error
- training set
- linear classifiers
- subspace information criterion
- generalization error bounds
- neural nets
- fault diagnosis
- supervised learning
- fuzzy logic
- back propagation
- learning machines
- perceptron learning
- genetic algorithm
- multilayer perceptron
- learning problems
- learning process
- artificial neural networks